According to several official European Union reports on the state of SMEs in the industrial sector, it appears that the ugliest of the key indicators of factory performance, overall equipment efficiency, has fallen in some countries to 60-70%, due mainly to shutdowns and malfunctions that reduce the availability of machinery. In addition, the direct costs associated with parts replacement and hours of manpower, particularly in mass production chains or those that work with high value raw materials, can have a huge impact on companies.
Consequently, the reliability of industrial plants, that is, the probability associated with correct operation during a specific period under normal conditions, is the baddy that must be tackled in an industrial outlook that is increasingly competitive with lower profit margins. Plant reliability is associated with the reliability of the assets of which it is comprised. Therefore, the element with the highest rate of shutdowns and/or malfunctions is the weakest link in the chain. It affects the entire system and therefore has an impact on final production capacity.
Currently, monitoring of processes and the associated machinery enables factory staff to analyze information and make improvements in the operation of the plant. The capacity to react quickly prior to shutdowns or malfunctions promotes optimum process management, the identification of opportunities to improve operation, and better production indicators. Indeed, predictive maintenance is the goody in the film.
Therefore, one of the areas of work of the Center Innovation Electronics. Motion Control and Industrial Applications (MCIA UPC), which is a member of the CIT UPC, is the development of technological solutions, to introduce into the market smart diagnostic systems for the monitoring and preventive maintenance of industrial machinery. This technology covers developments in the areas of instrumentation, electronic systems, communications and algorithms for diagnosis and prognosis, and the results are hardware and software systems that can be configured for specific applications.
As a result of the Center’s experience in the design and control of electromechanical systems, analysis of patterns of operation, malfunctions and their effects, and our extensive knowledge in signal processing, mathematical description and programming, we can provide tailored, high value-added solutions for the industrial sector.
At MCIA, we consider that the paradigm of industrial maintenance is moving towards machinery that includes its own diagnostic system, which is integrated into the rest of the equipment and indicates the state of the plant continuously, in a centralized way. One of our latest developments in this area was part of a European public-private initiative, carried out with two other research centers and five industry representatives who are associated in different ways with the industrial production sector.
As part of this initiative, a monitoring system based on acoustic emissions has been developed for rotating machinery, for the predictive maintenance of gears, bearings, axles and mechanical couplings. It can even be used whilst the machinery is in operation, as the diagnostic method has a negligible effect on the speed of the machinery. The project has been developed over a two-year period and is called Mosycousis.
The main advantage of detection using acoustic emissions over traditional vibration techniques is that a better signal-to-noise ratio is available, which makes early detection of a fault or degradation possible.
The concept behind this initiative consisted in developing smart sensors for distribution around a plant, combined with an expert system as a user interface to display the information in a centralized way. This represents a significant improvement on most current maintenance strategies in small and medium-sized factories. Unlike most current solutions, the monitoring system does not only measure the corresponding physical quantity, but the sensor itself includes the signal processing required to identify the content of the signal and provide information that has already been interpreted to the user.
By connecting various resonant transducers (elements that convert an acoustic signal into an electric one), the noise pattern can be extracted from the items of machinery that are being monitored. The transducers are connected to the same low cost, low energy consumption electronics, for better integration into the industrial environment. They are in a separate module that can be placed in one of the control cabinets close to the machinery. The local processor, which acquires the signals, is also located in the cabinet, and a processor for extracting useful information and generating a diagnosis of the monitored points.
During normal operation of the sensor, the diagnostic information is sent sporadically by short bursts of radiofrequency to the expert system, which should be installed in a computer in a control room close to or in the plant. The expert system, on which all the information generated by the sensor network is displayed, will show critical points that have been monitored and require a more detailed inspection. The same program will analyze the information in depth and provide diagnostic and prognostic data, using an analysis of the pattern of evolution. The processors included in the sensor and in the expert system convey patterns of acoustic emissions by the level of degradation of the monitored mechanical elements, using algorithms based on artificial intelligence that help to avoid up to 90% of false alarms.
After detection of a fault and complete analysis of the signals by the expert system, maintenance staff can carry out the corresponding actions, which include lubrication, replacement of parts, continuous monitoring of components for which the alarm has been raised, and prioritization of planned shutdowns, according to the company’s maintenance policy. The concept can be applied to numerous industrial applications. Nevertheless, the information that is obtained can be used not only for early diagnosis of faults, but also for integrating this data with external information, and for carrying out adaptive control to extend the useful life of machines.
This initiative is a clear example of socio-economic interest at international level in increasing the competitiveness of the industrial sector. Actions and technological developments such as those undertaken in this project will facilitate progress towards a smart, more competitive industrial sector.
The resources required to devise technological solutions often represent the link between industrial needs and the design of equipment to meet this need. Consequently, the MCIA Center represents a medium that provides technological innovation for the industrial sector; innovation that integrates high added-value products into plants, and thus helps companies to increase their competitiveness and obtain a greater market share.
PhD. Luis Romeral
Director of Motion and Control Apllications (MCIA UPC)
CIT UPC Center Member